Proper logging and monitoring

[empeje]

We have three kinds of deployment

• Earth
• Nation
• Community

For earth and nation it is important to log everything so that we can revisit and debug later when we need it. I think log archive is the first step and next step is something like Graylog where we push logs to statsd and it will send to Graylog, and it will give us flexibility when we want to search some problem in production. Thats for the backend (Nginx + CouchDB), for the frontend we should emit the data somewhere like Datadog or Newrelic for later use when debugging a problem.

For the community, it is okay to have the log stored in raspberry pi and the log of client in console.log of the user so when they have a problem they can tell us what is the problem based on the console.log.

So, my point here is we need to be careful with our nation and earth.

Since we also employ docker, we also need to monitor the system-level metrics like CPU, RAM, etc.

[empeje]

What is your plan @dogi about this?

If there is a problem with planet how the user report the problem and told us the step they did until getting an unexpected behavior?

[dogi]

@empeje good to think about this now
on the other side there are pieces which are more important which we need to tackle first

[empeje]

got it, will add label pending